Effects of bottom-up versus top-down digital cognitive training in older adults: A randomized controlled trial

Background: Impairments in bottom-up perceptual processing have been associated to the age-related cognitive decline in older adults. Digital cognitive training focusing on bottom-up and/or top-down processes have been studied as a tool to remediate age-related cognitive decline. However, it remains uncertain which training type and order of application has the most effective and lasting effects on the cognition. Methods: One hundred and fifteen older adults were randomly assigned to 40h of bottom-up then top-down or top-down then bottom-up digital cognitive training or an active control group. We evaluated cognition at baseline, after 20h and 40h of training and at follow-up. Results: Global cognition improved, for the top down group, after 20h of training (F_(1,39) = 4.28, p = 0.045) and for all three groups after 40h: Control group (F_(1,23)=10.25, p=0.004), top down/bottom-up (F_(1,39)=7.72, p=0.008) and bottom-up/top-down (F_(1,42)=7.02, p=0.011). The bottom-up/ top-down training improved  global cognition five months after the end of training  (F_(1,3)=11.98, p=0.041). There were also improvements in the recall cognitive domain, after 20h of training, for the bottom-up group (F_(1,11)=4.87, p=0.029)  and, after 40h, for all three groups: bottom-up/top-down training group (F_(1,46)=96.26,p<0.001), the top-down/bottom-up training group (F_(1,39)= 18.21, p<000,1) and the control group (F_(1,24)= 7.42, p=0.01). Processing speed was significantly slower after 40h for the control group (F(1,24)=4.45, p=0.04). Emotion recognition improved, after 20h, for the control group as compared to the therapeutic groups (F_(2,108) = 4.56, p=0.012). Conclusions: These results indicate that the bottom-up /top-down training has the most endurable effects in cognition, which reveals the importance of the order of the exercises in digital cognitive training.